Article Preview
Top1. Introduction
Engineering optimization problems arise in varities of engineering applications. These optimizations problems were generally difficult or impossible to solve using conventional optimization methods. In the last three decades or so, there has been a growing interest in applying Evolutionary Algorithm (EA) to engineering optimization problems. The most widely used EA is Genetic Algorithms (GAs) (Goldberg, 1989). Although GAs provide a robust and powerful adaptive search mechanism, they have several drawbacks. Some of these drawbacks include the slow convergence of the algorithm when solving complex problems, the problem of “genetic drift” which prevents GAs from maintaining diversity in the population, and the difficulty to optimally select the genetic operators such as population size, crossover and mutation rates (Davis, 1996; Yao, Kharma, & Grogono, 2010).
In the last few years, Particle Swarm Optimization (PSO) which belongs to the family of Swarm Intelligence has also been proposed as an alternative to GAs (Kennedy et al., 2001; Abido, 2001; Venayagamoorthy, 2005).
Recently, a novel type of Evolutionary Algorithm called Population-Based Incremental Learning (PBIL) which belongs to the family of Estimation of Distribution Algorithms (EDA) has received increasing attention (Baluja, 1994; Baluja & Caruana, 1995; Green, 1996; Folly, 2006, Sheetekela & Folly, 2010; Mitra et al., 2009). PBIL is simpler and more effective than GAs. In addition, PBIL has less overhead than GAs (Baluja & Caruana, 1995; Green, 1996; Gosling, 2004). The main reason for this simplicity is that in PBIL, the crossover operator is abstracted away and the role of population is redefined (Baluja, 1994; Baluja & Caruana, 1995). Furthermore, PBIL works with a probability vector (PV) which controls the random bit strings generated by PBIL and is used to create other individuals through learning. Learning in PBIL consists of using the current probability vector (PV) to create N individuals. The best individual is used to update the probability vector, increasing the probability of producing solutions similar to the current best individuals (Baluja & Caruana, 1995; Green, 1996). It has been shown that PBIL outperforms standard GAs approaches on a variety of optimization problems including commonly used benchmark problems (Baluja, 1994; Baluja & Caruana, 1995).